There is an excel file of 18 stocks with each forecast errors of 2 periods. Moreover, I added another excel file 'results', which consists of the calculted mean squared errors (MSE). I want to know if the MSE of period 2 is significantly higher than the MSE of period 1. Therefore, I think it's best to use the Diebold-Mariano test.
I need someone who is able to test this based on the DM-test and fill in the results in the excel-file. I think there are a lot of programs where the test is directly availble (eviews, matlab, octave, ...).
First, I have 2 questions:
- What is the advantage of working with the modified DM test? Do you think it's better to work with this test?
- Is it allow to use the squared forecast errors instead of the forecast errors for the DM test? Because I compare the MSE, which is the average of the square of the forecast errors.